In recent years, we have seen generative AI rapidly move through different eras: chatbots, image generation, voice, video generation, and more. But Dr. Fei-Fei Li, a longtime AI pioneer and co-director of Stanford’s Human-Centered Artificial Intelligence (HAI) Institute, is defining what she believes is the next frontier: spatial intelligence, an emerging field that she believes is “the defining challenge of the next decade,” as she wrote on Substack post this week.
That’s why Li co-founded World Labs in 2024 — and raised $230 million last fall — to build world models, or generative artificial intelligence models that the company says can “perceive, generate, reason, and interact with a 3D world.” This week, World Labs released its first commercial product, Marblewhich allows users to generate their own downloadable 3D worlds based on text, image or video prompts. Li wrote that she predicts spatial intelligence will potentially transform sectors from storytelling and filmmaking to architecture, robotics and scientific discovery.
“We see it [the] the world model is just as big and exciting, if not bigger [than the previous eras],” said Li, who is also the company’s CEO Edge in an interview.
“Bringing 3D to life and understanding the richness of spatial and 3D elements is simply a whole new level beyond the baseline of most other single modes,” said Ben Mildenhall, co-founder of World Labs Edge. He added that for all-human teams, “Building worlds like this is a huge problem. It takes such a large team, so much software, and time and effort… Think of the radical change that can come if you enable people to build things much faster – to ideate, iterate, and edit things in a much tighter loop.”
Marble offers four subscription tiers: Free, which allows for up to four global generations; Standard ($20/month), which allows for as many as 12 generations and more editing options; Pro ($35/month), which allows 25 generations and commercial rights; and Max ($95 per month), which allows for a maximum of 75 generations, as well as everything the professional tier offers. Edge was able to generate an open-air castle with waterfalls and, in generations of other users, explore ruined structures reclaimed by nature and spherical houses reminiscent of The Hobbit. You could take a few steps in such environments before you basically hit a wall in the 3D generation, and in the non-free levels the downloads are compatible with tools like Unreal Engine and Unity. Mildenhall said he has seen some people willing to put in long hours of work be able to create “pretty large interiors” using marble.
Mildenhall said he could imagine authors using it to build their imaginary world, or people working on visual effects or location searches in the film industry. At the enterprise level, he said he could imagine companies using Marble or one of World Labs’ future products to analyze and visualize enormous swaths of data.
“Even given the limitations of this model, in some emerging behaviors we see light beyond where we are,” Li said, adding that people can connect spaces in ways “beyond human imagination.”